Ecacy in Reduction of Lung Lesions of a Toxin Expressing Whole-cell Vaccine Against Multiple Serovars of Actinobacillus Pleuropneumoniae in Growing Pigs

Background: Actinobacillus pleuropneumoniae is a major economically signicant bacterial respiratory pathogen of pigs, and vaccine use is considered an integral control method to prevent disease. The objective of this multi-study analysis was to evaluate the serovar independent ecacy in growing pigs of the C-vaccine (Coglapix®, Ceva, France), which comprises whole cells of A. pleuropneumoniae serovars 1 and 2 expressing ApxI, ApxII and ApxIII toxins. Ecacy was based on protection against lung lesions, since there is good correlation between the severity/extension of lung lesions and losses induced by pleuropneumonia. Vaccine ecacy was determined against challenge with the most common serovars (1, 2, 4, 5, 6, 7, 9/11 and 13) of A. pleuropneumoniae in a total of 13 studies of the same design and reproducibility was validated. Results: Protection against homologous serovars 1 and 2 signicantly reduced lung lesion scores (LLS) compared to the positive controls: p = 0.00007 and p = 0.00124, respectively. The protection against heterologous serovars 4, 5, 6, 7, 9/11, and 13 also signicantly reduced LLS: range p = 2.9e-10 to p = 0.00953. Reproducibility between challenge studies was excellent with the estimated random effect of study (the fraction of the total variation attributed by differences between studies) being only 2.6%, 2.2% and 4.8% for three serovar 2, two serovar 9/11 and serovar 4 challenge studies, respectively. An outlier was the 35% of variation attributable to trial between the two serovar 6 challenges, possibly explained by a Streptococcus spp. outbreak. Conclusions: A highly signicant serovar independent reduction of pathological lung lesions by the C-vaccine was demonstrated for all the serovars tested (1, 2, 4, 5, 6, 7, 9/11 and 13). High levels of protection with similar signicance-values were obtained for both homologous and heterologous

such as caused by A. pleuropneumoniae, at the farm level 14,15,16,17 . To investigate pleuropneumonia in all its possible manifestations, pathological evaluation of lung lesions appears to be the least biased method. Performing this evaluation close to pneumonic infection would seem to reveal the most accurate validation of the degree of pleuropneumonic impact on the individual pig.
So far nineteen A. pleuropneumoniae serovars have been classi ed worldwide 41 . However, as the difference between serovar 9 and 11 is only one amino acid in the complete CPS loci and they have identical toxin pro les (ApxI, ApxII), these two serovars can be considered as one: serovar 9/11 42 . Strains belonging to different serovars are highly different in virulence and in some cases different strains of the same serovars can express different pathogenicity features; usually due to different Apx toxin pro les investigation 35,36,37,38,43 . More extrinsic factors like general stress 3,43,44 , poor air quality and climatic control, particularly high ambient temperature variations over the day, are associated with increased severity of pathological lesions caused by A. pleuropneumoniae, even in the case of what are considered low virulent A. pleuropneumoniae strains 3, 6 .
A. pleuropneumoniae has several virulence factors, some are well described, and several are under investigation. The three exotoxins: ApxI-III and lipopolysaccharide (LPS) have been proven responsible for lung lesions, but at the same time are both immunogenic and can induce protective immunity 1, 3, 46 . Many A. pleuropneumoniae virulence factors have been described 46 including outer membrane proteins (OMPs), some of which are immunogenic and therefore potential vaccine candidates 47 .
The variation in virulence between A. pleuropneumoniae serovars is mainly determined by the production of one or two of the ApxI-III toxins. These exotoxins providing nutrients for further growth and activity via lysis of the nearby cells in the lung tissue including neutrophils and macrophages 48,49 . LPS are both adhesion factors, allowing for colonisation and the production of exotoxins, and at the same time enhancing the cytotoxic effects of ApxI-III 50,51,52,53 .
Several commercial vaccines are available which differ in their composition and can be appointed into one of three A. pleuropneumoniae vaccine categories: 1) killed A. pleuropneumoniae whole-cell components only (bacterins); 2) subunit vaccines containing ApxI-III toxins only; and 3) a combination of these 47 . With distinct differences in e ciency, they all reduce clinical signs, but none can fully prevent infection and colonization 54 . Antibodies against ApxI-III are responsible for the serovar-independent protection against lung lesions 3, 6, 46, 47 . Due to limited cross protection between the serovars, bacterin vaccines lack e cacy compared to ApxI-III combined bacterin vaccines; pure toxoids vaccines lack in general protective capacity due to lack of LPS and other cell wall components 55,56,57,58 . A. pleuropneumoniae vaccine group 3 above is quite heterogenous, with a wide variety in composition, from a subunit vaccine containing the ApxI-III and only one of the cell wall OMPs, to the vaccine evaluated in this study based on whole-cell components of two serovars, together expressing all three of the ApxI-III toxins, and to vaccines containing whole-cell components of several A. pleuropneumoniae serovars together with some exotoxins.
A combination of the three exotoxins, ApxI-III with LPS, and likely more of the abundant cell-wall based antigens 56, 58 , induces a strong and speci c cell mediated immune response that can confer serovar independent protection 3 . This is an effective design for an e cacious serovar-independent vaccine, feasible for A. pleuropneumoniae prophylaxis to: increase animal well-being, reduce antimicrobial use, and reduce losses due to pleuropneumonia in all its manifestations at any A. pleuropneumoniae-endemic farm at any time 3,6,47 .
The objective of this study was to evaluate the e cacy of a vaccine comprising whole cells of A. pleuropneumoniae serovars 1 and 2 which in combination express ApxI, ApxII and ApxIII to protect against pleuropneumonic lungs lesions following challenge with multiple prominent serovars of A. pleuropneumoniae in growing pigs.

Materials & Methods
Data from thirteen studies each including one of the eight A. pleuropneumoniae serovars 1, 2, 4, 5, 6, 7, 9/11 and 13 performed over the period of 2011 to 2020 were available (Table 1). Where data on multiple studies with the same serovar were available weighted lung lesion score (LLS) of the vaccinated group of pigs (Vac) versus the non-vaccinated pigs in the positive control group (Pos) were pooled and analysed while taking the potential effect of study into account (Table 1). Also, variance between studies on the same serovars were analysed to estimate quality of repeatability.

The vaccine
The vaccine tested was Coglapix® (Ceva Santé Animale, France) hereafter referred to as C-vaccine. The Cvaccine is based on whole cells of A. pleuropneumoniae serovars 1 and 2 expressing ApxI, ApxII and ApxIII. Over the span of years, the vaccine composition and quality control has not changed. Apart from the cross-protective Apx-toxins, this vaccine contains all principal cell wall structures of A. pleuropneumoniae bacteria in undetermined quantities which contribute to A. pleuropneumoniaeprotective immune response: LPS, OMP's and the several other cell wall components.

Serotyping of the challenge strains
Strains belonging to different serovars were isolated from clinical cases of swine pleuropneumonia and serotyped using hyper immune sera by indirect haemagglutination as described previously 60 .
Serotyping of all A. pleuropneumoniae strains was con rmed in a multiplex-PCR based on capsular loci carried out as described by Bossé and colleagues 61 .

Calibration and preparation of the challenge strains
Strains were assessed for their ability of growth in liquid culture in a Tryptic soy broth supplemented with yeast extract and nicotinamide adenine dinucleotide solution in shake-asks rotated at 180 rpm and kept at 37 0 C. Their growth curve was analysed using sampling at pre-determined sampling points and subsequent optical density (OD) measuring at different wavelengths using a standard laboratory photometer. At each sampling point, the cultures were subjected to colony forming unit (CFU) counts using standard bacteriological techniques. The OD and CFU values were then aligned and the strainspeci c, optimal wavelengths were determined. After this initial procedure, in each case when a challenge trial was performed, the strain used was prepared in shake asks under regular OD monitoring and stopped when reaching the desired live titre based on the OD-CFU calibration curve.

Aerosol dosing technique
Challenge strains were propagated and used for the test when 10 9 CFU/ml concentration was reached.
The A. pleuropneumoniae stock was diluted in sterile PBS to achieve the optimal required 10 6 , 10 7 or 10 8 CFU/animal treatment dosage as shown in Table 1. Actual calculations were made at the test site, using the following parameters to introduce 1 dose/animal during the aerosol treatment to the chamber: Pig body weight 62 and volume Number of pigs placed in the chamber for one run (6, 8 or 10) Volume of chamber.
Volume of liquid, turned to aerosol by the ultrasonic nebulizer in 10 minutes (usually 100-150 ml, depending on air temperature and humidity) The pigs were evenly distributed and secured in the chamber by partition fences; aerosol was created by an ultrasonic humidi er and uniformly dispersed by internal ventilation. After 10 minutes of treatment, the pigs were kept in the chamber for an additional 2 minutes with the nebulizer switched off, to allow complete uptake of the aerosol droplets (fresh air was provided during this time to allow normal breathing). Before the rst run, the chamber was moisturized with the nebulizer to prevent aerosol loss caused by adherence to the dry surfaces; before each run piglets introduced to the chamber are given a couple of minutes of ease to ensure normal respiration before the doors are closed and the challenge is initiated. This aerosol chamber concept is developed by Palya and Kiss, and inspired by previous work on aerosol chambers 18, 32, 63 .

Intranasal (IN) challenge
Production of the challenge strain and calibration of the challenge dose was as described above. The cultures were prepared in 10× concentration of the desired challenge titre and diluted in sterile PBS to reach the working concentration. Each animal received 5 ml of challenge dose into each nostril using intranasal cannulas; the exact individual animal dose is shown in Table 1.

Pre-trial determination of challenge dose
Prior to using the strains in vaccine challenge trials, challenge dose calibration studies were performed. In these trials, three groups of 10 non-vaccinated A. pleuropneumoniae-negative pigs were challenged with doses of 10 6 , 10 7 or 10 8 CFUs, monitored daily for clinical signs and euthanized one week later. Mortality and LLS were evaluated to select the optimal challenge dose to be used: a mortality of 20-30 % in the non-vaccinated control group.
For A. pleuropneumoniae 2 and A. pleuropneumoniae 9/11 the pleuropneumonic impact of different concentrations of challenge dose were investigated via LLS to evaluate the A. pleuropneumoniae protective capabilities of the vaccine for different challenge loads.

Inclusion criteria
Pigs of either sex and of different breeds changing over time (Table 1) were recruited from farms free of A. pleuropneumoniae, Mycoplasma hyopneumoniae, toxin-positive Pasteurella multocida (progressive atrophic rhinitis), porcine reproductive and respiratory syndrome virus, Aujeszky's disease virus, classical swine fever virus and African swine fever virus based on regular PCR and/or serology tests performed either by government or private labs. Also, the animals had no previous clinical history of infection by Streptococcus suis and Glaesserella parasuis.
Any animal selected for inclusion in A. pleuropneumoniae challenge trials had to be negative in the ApxIV ELISA (IDEXX APP-ApxIV Ab) test when serum sampled at 5-6 weeks of age, con rming that they were negative for both A. pleuropneumoniae infection and colostral A. pleuropneumoniae antibodies.

Trial design
The challenge trials were performed using the same overall study design. The only difference was the use of an aerosol chamber challenge model by SSIU and an intra-nasal (IN) application by R&D (Table 1).
Detailed methods are given below.
Pigs at the age of 7-8 weeks, were randomly assigned to either non-vaccinated or vaccinated groups and At D42 all pigs individually received pre-determined equal doses of the relevant virulent A. pleuropneumoniae strains either by application in an aerosol chamber, or by the IN route, as described above.
At D49, one-week post-challenge, the trials were terminated. All live pigs were humanely euthanized and pathoanatomically evaluated to establish the individual lung lobe lesions to calculate the individual LLS.
Persons performing the pathoanatomical evaluation were not involved in vaccination and unaware of which pigs belonged to which test-groups.
For A. pleuropneumoniae 2, three studies, for A. pleuropneumoniae 4, 6 and 9/11, two studies, and for the remaining A. pleuropneumoniae serovars, one study were included in the analyses.

Post-mortem evaluation of weighted lung lesion score (LLS)
In the vaccination-challenge trials, all animals euthanised on day 7 post-challenge, D49, were subjected to necropsy to investigate the pathological changes due to actinobacillosis. Evaluation of the post-mortem lesions in the lungs and on the pleura were performed blind and in accordance with a previously described scoring system 15 . All seven lobes of the lung of each pig in trial were examined and each lobe scored on prevalence of pathological lesions of pneumonia and/or pleuritis (pleuropneumonia). Score valuing was according to the size of the affected area: absence = score 0, 1-20% = score 1, 21-40% = score 2, 41-60% = score 3, 61-80% = score 4, and 81-100% = score 5 15 .
Weighting This way each pig lung ended up with a total LLS of 0-5; the more lesions the higher the score.

Statistical analyses
The effect of vaccine on LLS was analysed using linear (mixed) models separately for each A. pleuropneumoniae serovar. If more than one study was available, a random effect of study to account for the possible clustering of effects within a study was included. To assess the importance of the between study variation, the intraclass correlation coe cient (ICC) was calculated as the proportion of the total variation attributed to the random effect. For the outcome (LLS), a limit of detection (LOD) was de ned as half the minimum observed LLS. The LOD was added to all LLS before it was log transformed to improve the underlying assumption about normal distribution of data. All analyses were done in R 65 , using the lme4 66 package for statistical analyses of mixed effects models with the lmerTest 67 package for testing of signi cant effects.

Results
The protection of the C-vaccine against the homologous A. pleuropneumoniae serovars 1 and 2 strains was demonstrated with highly signi cant reductions of LLS compared to the positive controls: p = 0.00007 and p = 0.00124 respectively ( Table 2). The protection of the C-vaccine against the heterologous A. pleuropneumoniae serovars 4, 5, 6, 7, 9/11, and 13 was demonstrated with equally highly signi cant reductions in LLS: p = 2.9e-10 to p = 0.00953 ( Table 2). The negative controls in all studies stayed A. pleuropneumoniae negative and without any LLS to be observed.

Discussion
The concept of combining the ApxI, ApxII and ApxIII 3, 6, 46, 47 with cellular components of A.
pleuropneumoniae 55,57 has been demonstrated to result in a highly effective serovar-independent vaccine that reduces lung lesions and mortality, and improves production performance. The use of such an e cacious vaccine will increase animal well-being, and reduce both antimicrobial use and economic losses due to endemic pleuropneumonia 3, 6, 47 .
To our knowledge, this is the most exhaustive testing on any A. pleuropneumoniae-vaccine, whether approved to be serovar independent or not by the relevant authorities. In these studies, we have analysed the e cacy of the C-vaccine in protecting against lung lesions of eight different virulent eld strains of A. pleuropneumoniae (1, 2, 4, 5, 6, 7, 9/11 and 13); six heterologous and two homologous with the serovars on which the vaccine is based. We found a signi cant reduction in LLS for the vaccinated groups compared to the positive controls. This implies that the vaccine is capable of inducing serovarindependent protection and is a valuable characteristic for optimizing the control of A. pleuropneumoniaerelated pig health problems.
The majority of the studies we carried out used an aerosol chamber. When considering challenge models, for most investigators, the choice is between IN or aerosol chamber challenge (AC). IN has an inherited accuracy in applied dose but is labour intensive and comparatively more expensive. In addition, dependent on pig handling and dose application (e.g., sedation/non-sedation), IN is potentially more stressful which can increase respiratory rate, hence respiratory volume and can affect the planned dose. With AC, skilful pig handling can ensure acceptance of the animals to the chamber and less stress. Our results indicate that reproducible protection studies can be performed using aerosol chambers with, to our knowledge, the Ceva Phylaxia aerosol chamber A. pleuropneumoniae challenge concept, being the only one validated for reproducibility using the intraclass correlation coe cient (ICC). Where determined, reproducibility between challenge studies was high, hence the outcome-of-trial data can be considered of high reliability, re ecting accurate individual challenge dose calculations, and basing the lung lesion scoring on standardized methodology 15 adapted to the biological appearance of the lung 64 achieving standardized, reproducible, weighted lung lesion score (LLS).
The variations attributable to differences between studies are very low: 2.6%, 2.2% and 4.8% for three A. pleuropneumoniae 2, two A. pleuropneumoniae 9/11 and two A. pleuropneumoniae 4 challenge studies, respectively. An outlier is the 35% of variation attributable to trial between the two A. pleuropneumoniae 6 challenges. An explanation could be that pleuritis was generally observed in a larger proportion of the animals in the 2019-study. Bacteriology demonstrated the presence of Streptococcus spp. in these samples. Signi cant improvement in LLS compared to the control group was still observed in this trial alone, and even more so when analysed together with the 2018-study. That infection with other pathogens, e.g., Bordetella pertussis, can affect lesion score in A. pleuropneumoniae challenged animals has been documented by others 30 .
Others have used aerosol chambers to challenge pigs with A. pleuropneumoniae 18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33 . In these sixteen publications ve common and important serovars were investigated: seven on serovar 1 19,20,21,22,31,32,33 , four on serovar 7 26, 28, 29, 30 , two on serovar 2 23, 24 , one on serovar 5 only 25 , one on serovars 2, 5 and 6 18 , and one on serovars 2 and 9 27 . Serovars 2, 5 and 6 were concluded as being of moderate to high virulence 18 , but this was based on small numbers of animals being investigated, and the result with serovar 5 can be considered surprising given that this is normally considered as of high virulence 48 . Serovar 7 was considered as moderately virulent 30 . Based on very high doses in identical trial designs, serovars 1 and 5 appear comparable in virulence measured on mortality only 20,25 . When comparing dosage and outcome empirically across the heterogenous trial designs, serovar 1 stands out as the most virulent closely followed by serovars 5 and 9, placing serovars 2, 6 and 7 as moderate to highly virulent. However, most of these studies, like ours, were not designed to reveal differences in virulence, rather dosing was aimed at obtaining similar disease severity distributions in the positive control groups to enable assessment of vaccine protection. Nonetheless, our data indicates broad agreement with the literature in that serovars 1, 5, 9/11 are the most virulent, serovars 2 and 13 of slightly lesser virulence, and serovars 4, 6 and 7 as moderate-to-highly virulent. It should be noted that the serovar 2 isolate we used was from Europe which expresses ApxII and ApxIII which is of higher virulence than serovars 2 isolates from North America which typically only express ApxII 37
To our knowledge the aerosol chamber challenge concept is the only one validated for reproducibility, basing challenge on individual pig challenge doses and weighted lung lesions scores for the most accurate biologically evaluation of disease and vaccine protection against A. pleuropneumoniae disease.

Declarations
Ethics approval and consent to participate The trial designs were all the same and in accordance with the European Pharmacopeia 59 . All trials were performed by Ceva Research and Development (R&D) Department or Ceva Scienti c Support and Innovation Unit (SSIU) in Hungary.
All studies followed local law and regulations. Authorization was provided by the Government O ce of Baranya County Food Chain-Safety and Animal Health Department, Hungary. Individual study approval ID noted ( Table 1). The

Consent for publication
Not applicable Availability of data and materials The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.

Funding
Ceva Santé Animal, France is the owner of Ceva Phylaxia SSIU and R&D departments and funding all costs relevant to these studies. All authors are employed by Ceva Santé Animale, however allocated to and paid by different departments. This except the statistician Dr Nils Toft, these speci c services paid by Ceva but accountable to own conclusions only.
Authors' contributions PM generated this multi-analysis, gathered the data and acted as primary, corresponding author. NT analysed all data in the capacity of being a skilled independent statistician experienced in multi and meta-analyses. IK together with VP developed the Ceva aerosol chamber concept and are responsible for the production of the challenge data of the SSIU, plus reviewed and contributed importantly in writing. HS reviewed and contributed importantly in writing. MT is responsible for the production of the challenge data of the R&D, plus reviewed and contributed importantly in writing.